How to create data source entities

Use data source entities in Bold360 AI to turn your data into smart conversations.

Make sure your widget is set to conversational. Learn more here.
In this example, we have a list of mobile phones and we want to create a bot that recommends users the best mobile devices based on their needs. We are going to set up a data source entity based on this list of mobile phones and then create an article that uses this entity.
  1. Create the list of mobile phones in a CSV file.

    A sample CSV file is attached at the end of this article. Learn more about the different data source options and requirements in this article.

  2. In the AI Console, go to Knowledge > Entities.
  3. Choose (Create Entity).
  4. Select data source entity for entity type.
  5. Provide a name and optionally a description for the entity.

    The entity name can contain letters, numbers and underscores. Spaces are not allowed.

    You can use the description to give the entity a more meaningful name or description that helps you identify it later.

    Example: Name your entity MOBILE_PHONES.

  6. Fill out the Data Source URL field.

    You can select the CSV file from your local computer and upload it to Bold360 AI' secure server or you can upload the file to a file hosting service like OneDrive and enter its URL here.

    Note: If you provide a link to the file's location, make sure it's publicly accessible.
    See What are the data source URL options? for more options.
  7. Set the response type to CSV.
  8. Choose LOAD.

    Result: The properties table is now loaded. Properties are the columns that are available in the data source file.

  9. Fill out the Properties table.

    See What are the property options for data source entities? and How do life cycle options work? to learn more about the options.

    Example:

    You should decide what questions the bot should ask the end users to offer the best possible recommendation about phones. In this example, the bot asks questions about brand, price range, and camera quality.

  10. Optional: Define fallback messages.

    See What are fallback messages? to learn more.

    Example: For this example, we enter Unfortunately, there's no mobile device that matches your preferences. for no results and set up a carousel to be displayed with the image, the brand and model, and the price if there are multiple results based on the customer's responses.

  11. Save your changes.
  12. Now go to Knowledge > Articles.
  13. Create a new article or open an existing one and add the entity in the article's body.

    In the article's body, create the final response of the bot: this is what the bot responds when there is a single mobile device to recommend. The response should use the entity to present the details of the selected mobile device. To insert an entity, you can either choose or enter [[ in the text editor and have a drop-down of all entities displayed. Once you add an entity to the response, the entity is automatically added to the article. You can expand the entity to see the questions that the bot may ask the user and adjust them to fit the article if needed.

    Note: You can drag-and-drop entities in the Article Editor to control the order of the questions in the conversation.

    Example: Intent: I want to buy a mobile phone

    Body: ?

  14. Publish the article.

You can now see the conversation in action in your widget ? just enter I want to buy a mobile phone.

Important: The bot doesn't always ask all the questions we set in the entity. It can automatically identify when there is no need to ask a question, for example, when there is only one possible result left. In addition, the order of the questions might differ as the bot automatically chooses the order based on what it estimates as the fastest way to get to a possible answer.